Cracketology: Predicting the Top-16 Reveal

Photo from CBS Sports broadcast

On Saturday, the NCAA will reveal their initial Top-16 seeds ahead of the NCAA Tournament. This always occurs after the Super Bowl and shows the exact order of the first four seed lines if the Tournament were to begin that day. Well...sort of. In each of the past two seasons, the Cracked Sidewalks projection from the beginning of the week was more accurate than the projection we released on the eve of the Reveal.

What this tells us is that when the Selection Committee gets together on Tuesday, they are using the metrics and results including games on Sunday or Monday. While it seems counter-intuitive, it makes sense. They likely start work on putting together the top four lines on Tuesday, so they continue to work with the same data all week. It's not dissimilar to the week of Selection Sunday. In the past, we've dug into the numbers that show Championship Week results have minimal impact on the Sunday bracket, so why would the week of the Top-16 Reveal be any different?

We are using the records, Quadrant results, and metrics from Tuesday morning because it seems unlikely anything that happened after that will impact what we see on Saturday. Today, we will go seed by seed through the Top-16 to explain not just who you can expect to see in the Reveal, but why they are there. We'll also include at the end some other teams that may be considered.

As far as things to consider, to fit a seed line teams should have average resume metric averages that are four times the seed line. This means being at an average of 4.0 or under for 1-seeds, 8.0 or under for 2-seeds, 12.0 or under for 3-seeds, and 16.0 or under for 4-seeds. Volume of wins in Quadrant 1A, Quadrant 1, and Quadrants 1+2 are also heavily considered, usually more so than losses. Having winning Quadrant records is also a positive.

1-Seeds

1-Auburn: The Tigers 21-2 record is better than anyone else in contention for the top lines. Their 11 Q1 wins and 15 Q1+2 wins are both most in the country. They rank #1 in 5 of the 7 metrics on the team sheet. This one is an easy choice.

2-Alabama: The Tide edge out Duke because they have the same overall record, 7 Q1 wins to 5, and 14 Q1+2 wins to 9. The teams are near metric mirrors, with Alabama averaging 4.0 in the metrics while Duke is at 3.8, but the strength of what Alabama has had to do in the SEC to match Duke's overall record is the deciding factor.

3-Duke: This is the last easy selection. While other teams have a few more Q1 wins, Duke is the only team left that has 5 of their 7 metrics ranking in the top-3.

4-Houston: This would be the first surprise of the bracket, and it will depend on what data the Committee is using. You could also argue Tennessee or Florida for this spot, but Houston has the same number of Q1+2 wins as Tennessee, more Q1 wins than Florida, and a better predictive average (2.0) than the Vols (4.3) or Gators (6.0). Ultimately, predictive metrics are thought to drive seeding, and combining that with the Cougars being the presumptive Big 12 champs (another factor that has weighed heavily) gives Houston the edge.

2-Seeds

5-Tennessee: Comparing the Vols to Florida, the Vols have more Q1 wins (8-3) and rank higher on five of the seven team sheet metrics, including a better NET, resume average, and predictive average.

6-Florida: If the Committee considers results through Tuesday, the Gators could jump up to the last 1-seed, but we're confident they won't be lower than this. Six of the seven team sheet metrics have the Gators in the top-6 and no one left has fewer losses (3) and more Q1+2 wins (11).

7-Purdue: If Purdue can improve their predictive average (9.0) they have a legitimate shot at a 1-seed. The Boilermakers' 7 Q1 wins trail only Auburn and Tennessee, while their 14 Q1+2 wins are tied with Alabama for second best in the country. Metrically, they are two points behind the six teams ahead of them, but three points better than anyone behind them. This is one where using later data could hurt as the Boilermakers took a loss to Michigan Tuesday night.

8-Texas A&M: Another tough call here because no one fits the mathematical 2-seed profile of 8.0 or better averages. The Aggies are metrically deadlocked (both 10.7 average) with Arizona but have more Q1+2 wins (12) than the Wildcats. Wisconsin has as many Q1+2 wins, but the Aggies are better in 5/7 team sheet metrics and have more Q1A wins (4) than Bucky (1). No other team can claim a metric edge nor better collection of wins than Buzz Williams' team, so the Aggies land here.

3-Seeds

9-Wisconsin: The strength of Wisconsin's resume are 7 Q1 and 12 Q1+2 wins. While their metric average (13.0) fits a 4-seed more than a 3-seed, they have better results than the only two teams that meet a 3-seed mathematical profile: Arizona and Iowa State.

10-Iowa State: Iowa State's 12.2 metric average is better than anyone left except for Arizona, who we'll get to later. The only competition there is Kansas (13.3) but ISU also has more Q1 wins (6 to 5) which keeps the Cyclones on this line.

11-Kentucky: Metrically, Kentucky should be a 5-seed. So why are they here? Quite simply, the wins. Their 6 Q1A wins are more than any other team left and behind only Auburn. Their three wins over top-5 teams (up to 4 on Tuesday) are more than any other team in the country, Auburn included. No one can match their quantity of quality, which bumps them up higher than their metrics would indicate.

12-St. John's: There are two teams that have a better metric claim to this spot. Arizona is at 10.7, Kansas is at 13.3, and St. John's is only at 15.2. The Wildcats (7) and Jayhawks (5) have more Q1 wins than St. John's (3), but the Johnnies only have 3 losses with both of their competitors having double that. The real differentiator is that St. John's is recognized as our Big East champion.  We've seen the Selection Committee nod to this in the past when Marquette was given the #8 overall in 2023.

4-Seeds

13-Arizona: The Wildcats fit a 3-seed perfectly. Their metric average is 10.7, no one left has more Q1 wins (7) than they do, and they don't have any bad losses. However, they also have more total losses than anyone ahead of them and they don't have any top tier wins away from home. It wouldn't surprise us to see them higher, but there just isn't room on the 3-line for them so a 4-seed feels appropriate.

14-Kansas: The Jayhawks edge Michigan State in 5 of the 7 team sheet metrics, a better metric average (13.3 to 15.0), have more Q1 wins (5 to 4), and more Q1+2 wins (10 to 9). They edge Texas Tech in 4 of 7 metrics with metric average edge (13.3 to 15.0 again), but have one fewer Q1 win (5 to 6). However, Kansas' Q1+2 advantage (10 to 7) and not having multiple Q2 losses kept them ahead.

15-Texas Tech: While MSU was a better comp for Kansas, Texas Tech gets the slight nod over Sparty. They are in a metric dead heat, though TTU is better in 4 of 7 resume metrics when NET is included. The Red Raiders also have more Q1 wins (6 to 4) and digging deeper, the win at Houston is something Sparty just can't match. If they factor the Tuesday games, it favors the Red Raiders even more.

16-Michigan State: Metrically, no one left is close to Sparty. Their 15.0 metric average is far better than the next closest team (Michigan at 18.5). There are seven other teams with metric averages below 24.0 and MSU. Illinois is the only one that is ahead of Michigan State in as many as three metrics, the rest Sparty leads in at least six of seven. The biggest question is which data set they use. If it is from Tuesday morning, Michigan State belongs here. If they include this week's results, the edge likely goes to Michigan who we have at #17. Which of the Michigan schools is a protected seed likely tells us which games they included.

In Consideration

17-Michigan: The Wolverines have better metrics than Kentucky and added the big Purdue win this week. 10 Q1+2 wins is excellent, but the two Q2 losses knocked them down here.

18-Marquette: The Golden Eagles boast 11 Q1+2 wins, but shoddy metrics (20.7 average) and three Q2 losses keep them out of the Top-16.

19-Ole Miss: This team fits the 5-line perfectly. Nine Q1+2 wins, a losing Q1 record, and 6 of 7 metrics outside the top-16 keep them outside the conversation.

24-Memphis: The Tigers are a wildcard here. Their resume average is 13.3 but all predictive metrics are sub-40. Their metric profile fits an 8-seed, but the Selection Committee has boosted conference champs with great W/L profiles before (2022 Providence). I don't think they'll be here, but they could be the out-of-the-blue shocker.

Here's the current S-Curve and bracket:

Multibid Leagues

SEC: 13

Big 10: 10

Big 12: 7

ACC: 5

Big East: 4

MWC: 3

WCC: 2